Calculate working time difference between dates

Question

I need to calculate the average time between the arrival of a task and its completion. But if the task arrives on Friday, the time becomes too long due to the fact that the task is not executed on the weekend:

  DT_APP_ENTERED  IN_WORK
0     2023-02-17    54.80
1     2023-02-20    10.36
2     2023-02-21    10.68
3     2023-02-26    19.87
4     2023-02-27    11.35
5     2023-02-28    10.68
6     2023-03-01    10.91
7     2023-03-02    14.36
8     2023-03-03    36.46

How can I correctly add days to a datetime if weekday is friday?

I've tried to make a function:

def add_days(start_datetime):
    start_datetime = month['DT_APP_ENTERED']
   
    if start_datetime.dt.weekday == 4:
        start_datetime + dt.timedelta(days=2)
    else:
        start_datetime
    return (start_datetime)

But when I've applied it to a DF, there became an error: Truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all()

Not just Fridays. How about Christmas? Holidays in general? You need a calendar with working days, you can't exclude weekends. — Panagiotis Kanavos, Mar 17 '23 at 09:43
Does this answer your question? [Number of days between 2 dates, excluding weekends](https://stackoverflow.com/questions/3615375/number-of-days-between-2-dates-excluding-weekends) — Panagiotis Kanavos, Mar 17 '23 at 09:45
The possibly duplicate question shows NumPy's [busday_count](https://numpy.org/doc/stable/reference/generated/numpy.busday_count.html) can be used for this and can accept custom calendars as [busdaycalendar](https://numpy.org/doc/stable/reference/generated/numpy.busdaycalendar.html) objects — Panagiotis Kanavos, Mar 17 '23 at 09:47
@Panagiotis Kanavos, I've exclude holidays from calculation np.busday_count, counts only dates, I need to calculate datetimes (if I think correctly) — Maxim Merkushin, Mar 17 '23 at 10:19
The code you posted doesn't. If you use `busday_count` you can specify extra holidays as a parameter — Panagiotis Kanavos, Mar 17 '23 at 10:22
@Corralien, yes, I'd like to etit date, as if task arrived on Sunday — Maxim Merkushin, Mar 17 '23 at 10:24

Corralien · Accepted Answer · 2023-03-17T09:59:02.377

I suppose you use pandas. You can use pd.offsets.BDay:

df['NEW_DATE'] = (df['DT_APP_ENTERED'].where(df['DT_APP_ENTERED'].dt.dayofweek==4)
                                      .add(pd.offsets.BDay(1))
                                      .fillna(df['DT_APP_ENTERED']))
print(df)

# Output
  DT_APP_ENTERED  IN_WORK   NEW_DATE
0     2023-02-17    54.80 2023-02-20  # Friday -> Monday
1     2023-02-20    10.36 2023-02-20
2     2023-02-21    10.68 2023-02-21
3     2023-02-26    19.87 2023-02-26
4     2023-02-27    11.35 2023-02-27
5     2023-02-28    10.68 2023-02-28
6     2023-03-01    10.91 2023-03-01
7     2023-03-02    14.36 2023-03-02
8     2023-03-03    36.46 2023-03-06  # Friday -> Monday

If you prefer, the new day is Sunday, replace pd.offsets.BDay(1) by pd.Timedelta('2D'):

df['NEW_DATE'] = (df['DT_APP_ENTERED'].where(df['DT_APP_ENTERED'].dt.dayofweek==4)
                                      .add(pd.Timedelta('2D'))
                                      .fillna(df['DT_APP_ENTERED']))
print(df)

# Output
  DT_APP_ENTERED  IN_WORK   NEW_DATE
0     2023-02-17    54.80 2023-02-19  # Friday -> Sunday
1     2023-02-20    10.36 2023-02-20
2     2023-02-21    10.68 2023-02-21
3     2023-02-26    19.87 2023-02-26
4     2023-02-27    11.35 2023-02-27
5     2023-02-28    10.68 2023-02-28
6     2023-03-01    10.91 2023-03-01
7     2023-03-02    14.36 2023-03-02
8     2023-03-03    36.46 2023-03-05  # Friday -> Sunday

Calculate working time difference between dates

1 Answers1